privacy-review Skill | Agent Skills

<skill> <name>privacy-review</name> <description> Reviews data handling practices to reduce privacy risk and support compliance. Maps data flows, identifies PII, validates retention, and recommends minimization. Runs before features collecting or processing personal data. </description> <when-to-use> <trigger>New feature collecting user data (forms, tracking, uploads).</trigger> <trigger>Changes to data storage, retention, or deletion logic.</trigger> <trigger>Third-party data sharing or integrations.</trigger> <trigger>Analytics, logging, or telemetry additions.</trigger> <trigger>New user-facing consent or preference features.</trigger> <trigger>Cross-border data transfer changes.</trigger> <scope>Run before implementation; revisit when data handling changes.</scope> </when-to-use> <data-classification> <category name="PII (Personal Identifiable Information)"> <examples>Name, email, phone, address, IP address, device ID, SSN, passport.</examples> <handling>Encrypt in transit and at rest; strict access controls; retention limits.</handling> <risk>High - regulatory impact, user harm if breached.</risk> </category> <category name="Sensitive PII"> <examples>Health data, financial data, biometrics, race, religion, sexual orientation.</examples> <handling>Highest protection; explicit consent; consider not storing.</handling> <risk>Critical - severe regulatory and reputational impact.</risk> </category> <category name="Behavioral Data"> <examples>Browsing history, click patterns, usage analytics, preferences.</examples> <handling>Anonymize where possible; clear consent; retention limits.</handling> <risk>Medium - can identify users when combined with other data.</risk> </category> <category name="Business Data"> <examples>Transactions, contracts, internal communications.</examples> <handling>Access controls; backup; retention per business needs.</handling> <risk>Medium-Low - competitive/legal risk, not privacy-focused.</risk> </category> </data-classification> <review-checklist> <category name="Data Inventory"> <check>What data is collected? (List all fields/attributes)</check> <check>What is the data classification? (PII, Sensitive, Behavioral, Business)</check> <check>Is each data element necessary? (Minimization check)</check> <check>Can any data be anonymized or pseudonymized?</check> </category> <category name="Collection"> <check>Is consent obtained before collection?</check> <check>Is the purpose clearly communicated to users?</check> <check>Is collection limited to stated purpose?</check> <check>Are there legitimate grounds for processing? (Consent, contract, legal, vital, public, legitimate interest)</check> </category> <category name="Storage"> <check>Where is data stored? (Region, provider, on-prem vs cloud)</check> <check>Is data encrypted at rest?</check> <check>Are backups encrypted and access-controlled?</check> <check>Is storage location compliant with data residency requirements?</check> </category> <category name="Access"> <check>Who can access the data? (Roles, services, third parties)</check> <check>Is access logged and auditable?</check> <check>Is principle of least privilege applied?</check> <check>Are there controls for privileged access?</check> </category> <category name="Retention"> <check>How long is data retained?</check> <check>Is retention period documented and justified?</check> <check>Is there automated deletion after retention period?</check> <check>Are backups included in retention policies?</check> </category> <category name="Deletion"> <check>Can users request data deletion?</check> <check>Is deletion complete (including backups, logs, caches)?</check> <check>Is deletion verified and auditable?</check> <check>How are deletion requests handled for shared data?</check> </category> <category name="Sharing"> <check>Is data shared with third parties?</check> <check>Are Data Processing Agreements in place?</check> <check>Is cross-border transfer compliant (SCCs, adequacy)?</check> <check>Can users opt out of sharing?</check> </category> </review-checklist> <data-flow-mapping> <step number="1">Identify data entry points (forms, APIs, imports, tracking).</step> <step number="2">Map data movement: entry → processing → storage → output.</step> <step number="3">Note each system/service that touches the data.</step> <step number="4">Identify data exits: exports, APIs, reports, third-party sharing.</step> <step number="5">Document retention and deletion points in flow.</step> <format>Use simple diagram or table: Source → Process → Storage → Access → Exit.</format> </data-flow-mapping> <output-format> <section name="Scope">Feature/component being reviewed.</section> <section name="Data Inventory">Table: Data Element | Classification | Purpose | Necessary?</section> <section name="Data Flow">Diagram or description of data movement.</section> <section name="Checklist Results">Pass/Fail/N-A per category with notes.</section> <section name="Risks Found">Privacy risks with severity and likelihood.</section> <section name="Recommendations">Minimization, anonymization, control improvements.</section> <section name="Compliance Notes">Relevant regulations (GDPR, CCPA, etc.) and status.</section> <section name="Actions">Required changes before release, assigned to developer.</section> </output-format> <anti-patterns> <anti-pattern>Collecting data "just in case" without specific purpose.</anti-pattern> <anti-pattern>No retention policy (data kept forever).</anti-pattern> <anti-pattern>Logging PII in application logs.</anti-pattern> <anti-pattern>Third-party sharing without DPA.</anti-pattern> <anti-pattern>Assuming anonymization is sufficient without verification.</anti-pattern> <anti-pattern>Privacy review after launch (too late).</anti-pattern> </anti-patterns> <integration> <with-skill name="threat-modeling">Run together; threat model covers attack vectors, privacy covers data handling.</with-skill> <with-skill name="release-readiness">High privacy risks may block release.</with-skill> <with-agent name="developer">Implements minimization and controls.</with-agent> <with-agent name="architect">Reviews data flow design decisions.</with-agent> </integration> <references> <reference>docs/data.md for data model details.</reference> <reference>docs/security.md for access control model.</reference> </references> </skill>

Agent Skills: privacy-review

Install this agent skill to your local

Skill Files